Refitting Estimated Glomerular Filtration Rate Equations to Find A Fit For All

In the United States, more than 37 million adults—or 1 in 7 people—are affected by kidney disease. In its earliest stages, kidney disease is silent, and approximately 90% of people who have chronic kidney disease (CKD) are unaware. Many remain undiagnosed until they have kidney failure. Kidney disease is associated with complications in every organ system, and more deaths from kidney disease are caused by cardiovascular disease than by kidney failure. Early detection of kidney disease is critical, because clinical and lifestyle interventions can slow disease progression and reduce complications.

Now, new guidance from AACC Academy and the National Kidney Foundation (NKF) will enable laboratory medicine experts and clinicians to work together to implement the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) 2021 equations. The guidance emphasizes improved detection in high-risk populations and marks a significant leap forward in emphasizing an inclusive approach for all populations.

This landmark guidance calls for collaborative efforts among clinical laboratories and providers to drive meaningful change. Multidisciplinary teams are urged to unite to achieve improved disease detection, particularly in high-risk populations.

The Evolution of Laboratory Testing for Chronic Kidney Disease

Kidney disease diagnosis and staging primarily relies on laboratory testing. Glomerular filtration rate (GFR) currently serves as the best overall index of kidney function, and albuminuria as the most widely studied marker of kidney damage. Direct measurement of GFR requires measuring the clearance of an exogenous filtration marker and is impractical in routine clinical practice.

However, GFR can be estimated (eGFR) by combining endogenous filtration markers, such as creatinine or cystatin C, with surrogates for non-GFR determinants that influence filtration marker concentrations independent of GFR. Including these surrogates in calculation of the eGFR improves its accuracy relative to measured GFR (mGFR).

Because clinical laboratories routinely measure creatinine as part of the basic or comprehensive metabolic panel, the routine reporting of eGFR from creatinine (eGFRcr) has transformed medical practice. Notably, the introduction of eGFR into clinical practice has resulted in significant advances in kidney disease care, including federal investments in kidney disease surveillance programs and standardization of laboratory assays for creatinine.

Over the past few years, eGFR equations have been scrutinized increasingly because of their inclusion of a race variable (African American vs. non-African American) in eGFRcr calculation. The most widely used eGFR_crequations—the 4-variable Modification of Diet in Renal Disease (MDRD) and CKD-EPI 2009 equations—incorporate the variables age, sex, and race in addition to creatinine concentration (1,2). Use of age, sex, and race as surrogates for the non-GFR determinants of creatinine increase the accuracy of eGFR_crcomputed by the MDRD and CKD-EPI 2009 equations.

During the derivation of these equations, investigators observed that the average creatinine concentration in the study participants who self-identified as African American (12% in the MDRD study and 31% in the CKD-EPI development and internal validation populations) was higher compared to non-Black participants of the same age, sex, and mGFR (1,2). Inclusion of a race variable in these equations provides a higher eGFR_crin African American persons compared to non-African American persons of the same age, sex, and creatinine concentration.

At the time, researchers hypothesized that all three variables—age, sex, and race—were related to creatinine generation by muscle mass or diet. Of interest, blood cystatin C concentrations differ less with age, sex, and race, and race was not included in the CKD-EPI 2012 equation using cystatin C (eGFR_cys). Race, however, was included in the CKD-EPI 2012 equation using both creatinine and cystatin C (eGFR_cr-cys) which is more accurate than either eGFR_cror eGFR_cys(3). eGFR_cysand eGFR_cr-cyswere recommended as confirmatory tests (4).

We now recognize that race is a social construct rather than a biological variable. Including race as a distinct variable in eGFR and other clinical algorithms incorrectly reinforces race as a biological construct, and it overlooks the social costs of systemic racism.

Race is a particularly problematic surrogate because it was historically co-opted for the purpose of oppression, and should not be confused with genetic similarity, which is measurable and can be used to infer ancestry (5). While population-level racial differences are observed in creatinine, the individuals that comprise these socially constructed racial groups are both genetically and economically diverse. Race can serve as a proxy for the unequal access to and allocation of healthcare resources, poorer quality of healthcare, and worse social determinants of health experienced by racialized people. Current guidelines recommend that the socioeconomic and environmental stressors that influence human health and disease should be directly measured (5). Race should not be used as a proxy.

Racial disparities are evident at every stage in the detection, evaluation, and management of kidney disease. Black and Hispanic people are less likely to receive patient-centric dialysis treatment, and they are more likely to develop end stage kidney disease. Black and Hispanic people are overrepresented in neighborhoods with high levels of socioeconomic deprivation, which is associated with poorer kidney health, less access to medications that delay kidney disease progression, and delayed nephrology referral (6,7).

For this reason, it was especially appropriate to reconsider the use of race in eGFR equations. At the time of writing, there is no conclusive evidence that inclusion of race in eGFR equations contributed to these disparities; however, a unifying approach with potential consequences that do not disproportionately affect any one group of individuals was needed. In addition, an approach that avoided typological thinking that assumes that people can be grouped into distinct, homogeneous categories by race was desirable.

In November of 2020, NKF and the American Society of Nephrology (ASN) formed a task force to reassess the inclusion of race in diagnosing kidney diseases. The task force subsequently recommended immediate implementation of the CKD-EPI 2021 eGFR_crequation that does not include a race variable (8,9). This recommendation was supported by the AACC eGFR and Race Equity Task Force, which conducted a systematic review of the use of race in the eGFR equations and found no evidence to support this practice (10).

The CKD-EPI 2021 eGFR_crequation was developed using the same data sources in the derivation of the 2009 CKD-EPI eGFR_crequation (9). The refit equation achieves the goal of removing race, and while it is less accurate in both Black and non-Black subgroups, it is sufficiently accurate for clinical use for both Black and non-Black people. Similar findings were observed for the CKD-EPI 2021 eGFR_cr-cyswithout race that was developed by refitting the coefficients in the 2012 development and internal validation population without race.

How the New Guidance Will Lead to Health Equity

In the July 2023 issue of the Journal of Applied Laboratory Medicine, the AACC/NKF Guidance Document on Improving Equity in Chronic Kidney Disease Care was published; it's a new to implement the Task Force recommendations. The guidance document provides recommendations on the implementation of cystatin C testing, since the NKF-ASN Task Force called for increased use of cystatin C, and describes clinical situations in which the CKD-EPI 2021 eGFR_cr-cysor 2012 eGFR_cysmay provide more accurate estimates than the CKD-EPI 2021 eGFR_crequation.

The implementation of race-agnostic eGFR equations by clinical laboratories represents important progress towards racial equity in kidney disease, and the field of nephrology is the first to course-correct the harmful practice of including race in clinical calculators. The guidance document highlights multiple other considerations for clinical laboratorians as we collaborate with our nephrology colleagues in the quest to achieve kidney disease equity.

As clinical laboratorians, we are intimately familiar with the multiple potential sources of error that can bias our quantitative measurements. Although eGFR represents a reasonably accurate estimate of “population average mGFR,” it is not a precise estimate of an individual’s mGFR (11). There are numerous sources of error in eGFR_cr(12). In addition to errors in the creatinine assay (more pronounced with the Jaffe than the enzymatic methods), there are errors due to excessive variation in non-GFR determinants of creatinine (extremes of muscle mass or diet and inhibition of tubular secretion of creatinine), as well as errors in mGFR, which inflate the magnitude of error observed when eGFR is compared to mGFR.

While nephrologists use eGFR_crwith nuance, non-nephrology providers often consider the reported eGFR_crto be an exact value with respect to true GFR. Furthermore, reliance on eGFR_cras the sole measure of kidney disease—without confirmation by eGFR_cysor eGFR_cr-cysand without assessment of albuminuria—can delay kidney disease detection, evaluation, and management. The clinical laboratory has a lot to offer in this regard, including using enzymatic creatinine assays instead of Jaffe methods, cystatin C testing, urine albumin-to-creatinine ratio (ACR), and combining ACR testing with eGFR into a single test order: the Kidney Profile.

It also is appropriate to consider the use of sex in eGFR equations. Unlike race, sex is a biological factor. However, the associations of sex with blood creatinine and cystatin C concentrations are variable, and in some circumstances, the harms from using a binary sex variable may not be worth the added accuracy in GFR estimation.

For example, applying a sex variable for transgender persons is inherently problematic since gender-affirming hormones shift muscle mass and fat distribution. Changes in creatinine concentration after initiation of gender-affirming hormones vary between sexes, and individuals and can be influenced by gender-affirming hormone dose and duration of use (13).

This is further complicated by conflicting data describing the impact of gender affirming hormones on creatinine concentrations. For example, some studies show that people on masculinizing or feminizing hormones experience modest increases or decreases in creatinine concentration over time, respectively, while others report that use of either exogenous hormone confers no change in creatinine measurements. The sex-identifier used to compute eGFR is institution-specific and can be based on the capability of the electronic health record in use.

Sex used to compute eGFR may be sex assigned at birth, legal sex, or gender identity. None of these, however, take into account the effects of gender-affirming hormone therapies and their short and long-term impact on endogenous filtration markers and other physiological variables that are relevant to kidney disease measures.

Thus, interpreting eGFR in transgender people requires a holistic approach that takes into consideration the individual’s muscle mass (for estimates that incorporate creatinine) and hormonal therapy. Most importantly, their gender identity must be respected. Of interest, eGFR_cys equations are available that were developed without including sex as a variable and do not require specification of sex for calculation of eGFR (14,15), but there is little experience in the U.S. with these equations.

A Call for Collaboration

Calculating eGFR has transformed the interdisciplinary relationship of nephrology and laboratory medicine—a collaboration that has undoubtedly improved care across populations. However, fundamental to this laboratory result is the little “e” in eGFR. This value is an estimate for all people but is only one of several available laboratory measures of kidney disease. Collaboration among laboratory medicine and nephrology specialists can improve the use of these measures for detection, evaluation, and management of kidney disease.

Finally, there is a broader lesson to be learned from resolving the controversy about using race in eGFR equations: Biological variability and socioeconomic and environmental stressors must be considered in clinical research and practice to achieve more equitable care.  

References

1. Levey AS, Bosch JP, Lewis JB, et al. A more accurate method to estimate glomerular filtration rate from serum creatinine: A new prediction equation. Ann Intern Med 1999; doi: 10.7326/0003-4819-130-6-199903160-00002.
2. Levey AS, Stevens LA, Schmid CH, et al. A new equation to estimate glomerular filtration rate. Ann Intern Med 2009; doi: 10.7326/0003-4819-150-9-200905050-00006.
3. Inker LA, Schmid CH, Tighiouart H, et al. Estimating glomerular filtration rate from serum creatinine and cystatin C. N Engl J Med 2012; doi: 10.1056/NEJMoa1114248.
4. Official Journal of the International Society of Nephrology KDIGO 2012 Clinical Practice Guideline for the Evaluation and Management of Chronic Kidney Disease.
5. National Academies of Sciences Engineering and Medicine. Using Population Descriptors in Genetics and Genomics Research: A New Framework for an Evolving Field. Washington DC; 2023.
6. Powe NR. To have and have not: Health and health care disparities in chronic kidney disease. Kidney Int 2003; doi: 10.1046/j.1523-1755.2003.00138.x.
7. 2022 USRDS Annual Data Report: Epidemiology of kidney disease in the United States. 2022.
8. Delgado C, Baweja M, Crews DC, Eneanya ND, Gadegbeku CA, Inker LA, et al. A Unifying Approach for GFR Estimation: Recommendations of the NKF-ASN Task Force on Reassessing the Inclusion of Race in Diagnosing Kidney Disease. J Am Soc Nephrol 2021; doi: 10.1681/ASN.2021070988.
9. Inker LA, Eneanya ND, Coresh J, et al. New creatinine- and cystatin C–based equations to estimate GFR without race. N Engl J Med 2021; doi: 10.1056/NEJMoa2102953.
10. Marzinke MA, Greene DN, Bossuyt PM, et al. Limited evidence for use of a black race modifier in eGFR calculations: A systematic review. Clin Chem. 2022; doi: 10.1093/clinchem/hvab279.
11. Shafi T, Zhu X, Lirette ST, et al. Quantifying individual-level inaccuracy in glomerular filtration rate estimation: A cross-sectional study. Ann Intern Med 2022; doi: 10.7326/M22-0610.
12. Inker LA, Levey AS. Knowing your GFR—when is the number not (exactly) the number? Kidney Int 2019; doi: 10.1016/j.kint.2019.03.026.
13. Krupka E, Curtis S, Ferguson T, et al. The effect of gender-affirming hormone therapy on measures of kidney function. Clin J Am Soc Nephrol 2022; doi: 10.2215/CJN.01890222.
14. Pottel H, Björk J, Rule AD, et al. Cystatin based equation to estimate GFR without the inclusion of race and sex. N Engl J Med 2023; doi: 10.1056/NEJMoa2203769.
15. Ottosson Frost C, Gille-Johnson P, Blomstrand E, et al. Cystatin C-based equations for estimating glomerular filtration rate do not require race or sex coefficients. Scand J Clin Lab Invest 2022; doi: 10.1080/00365513.2022.2031279.

Christina C. Pierre, PhD, DABCC, FADLM, is a clinical assistant professor in the department of pathology and laboratory medicine at the Perelman School of Medicine at the University of Pennsylvania in Philadelphia, and a clinical chemist and section director of clinical chemistry and coagulation testing at Penn Medicine Lancaster General Hospital in Lancaster, Pennsylvania. Email: [email protected]

Dina N. Greene, PhD, DABCC, FADLM, is an associate professor in the department of laboratory medicine and pathology at the University of Washington in Seattle, and an associate laboratory director at LetsGetChecked Laboratory in Monrovia, California. Email: [email protected]

Mark Marzinke, PhD, DABCC, FADLM, is a professor of pathology and a professor of medicine in the Division of Clinical Pharmacology at Johns Hopkins University School of Medicine in Baltimore. Email: [email protected]

Melanie Hoenig, MD, is an associate professor of medicine in the Renal Division at Beth Israel Deaconess Medical Center in Boston. Email: [email protected]. 

Andrew S. Levey, MD, is a professor of medicine and Dr. Gerald J. and Dorothy R. Friedman professor emeritus at Tufts University School of Medicine, and chief emeritus of the William B. Schwartz Division of Nephrology at Tufts Medical Center in Boston. Email: [email protected]